Regular
Discriminating From
Borderline Other Axis
Mary C. Frances
Personality II Disorders
phrenic, The authors used the Revised Diagnostic Interview f or Borderline Patients to assess 22 clinical features of 120 patients with borderline personality disorder and 1 03 control subjects with other axis II disorders. Four of the 22 features were common in but nondiscriminating for borderline disorder, 1 1 were discriminating f or but nonspecific to borderline disorder, and seven were more specific to borderline disorder. The authors conclude that many clinical features thought to be indicative of borderline disorder are better viewed as personality
disorder
traits
and
that
the seven
more
spe-
cific features, alone or in conjunction with one another, may be particularly useful markers for borderline personality disorder. (Am J Psychiatry 1990; 147:161-167)
N
umerous clinical
studies features
den (1-29). Taken lighted four areas characteristic
of
have
detailed
of borderline
together, these of phenomenology borderline
the characteristic personality
disor-
studies have highas particularly
patients:
1) intense,
dys-
phoric affect, 2) transient cognitive difficulties, particularly those of a paranoid and dissociative nature, 3) widespread impulsivity, particularly of a physically self-destructive nature, and 4) intense, unstable relationships. Numerous studies have also shown that the clinical features of borderline personality disorder can be successfully discriminated from those of schizo-
Received March 2, 1989; revision received July 12, 1989; accepted Aug. 3, 1989. From the Psychosocial Research Program, McLean Hospital, and the Department of Psychiatry, Harvard Mcdical School, Boston. Address reprint requests to Dr. Zanarini, Psy. chosocial Research Program, McLean Hospital, 115 Mill St., Belmont, MA 02178. Copyright C 1990 American Psychiatric Association.
Am
J
Psychiatry
147:2,
February
1990
Disorder
Ed.D., John G. Gunderson, M.D., M.D., and Deborah L. Chauncey,
Zanarini, Frankenburg,
R.
Articles
atnic
depressed,
control
A.B.
nonpsychotic,
subjects
(2-7,
and general
9-13,
16,
18,
20,
psychi26,
29).
However, the descriptive validity of borderline personality disorder will remain in question until it is proven that
borderline
from
patients
particularly
patients
with the
can
other
other
forms
disorders
be
reliably
distinguished
of personality in the
disorder,
“dramatic”
clus-
ten of DSM-III. Although numerous studies using control subjects with other axis II disorders have assessed the diagnostic efficiency of the DSM-III criteria set for borderline personality disorder (15, 22, 24, 25, 27, 28), only three studies have made any effort to determine if the clinical features of borderline patients were significantly different from those of patients with other forms of personality disorder (8, 14, 17). One of these studies, by Sheehy et al. (8), compared the checklist ratings given to 45 outpatients clinically diagnosed as having borderline personality disorder with those given to a control group of 30 outpatients clinically
diagnosed
as
having
some
other
type
of
DSM-II personality disorder. Using a discniminant function analysis, they found that borderline patients had a higher score than the control subjects on nine of the
16 items
pulses, periodic chronic
studied:
unusually
strong
feelings
and
im-
trouble tolerating dysphonic affect, anhedonia, paranoid distortions of reality, an episodic on problem
with
impulse
control,
a history
of
promiscuity, episodic abuse of alcohol and/on drugs, a tendency to vacillate between idealizing and devaluing others, and difficulty maintaining a stable sense of self in a variety of external situations and internal mood states. In the second study, Barnash et al. (14) administered the Diagnostic Interview for Borderline Patients (DIB) (30) to 18 inpatients with clinical diagnoses of borderline personality disorder and to a control group of 31 inpatients with clinical diagnoses of some other type of
161
DISCRIMINATING
BORDERLINE
disorder. statements
personality
summary mated
two
types
PERSONALITY
DISORDER
Using cluster analysis of the 29 of the DIB, these authors discrim-
of borderline
patients
and
found
that
both types of patients were significantly more likely than the control subjects to report three clinical featunes: depressive affect, self-mutilation, and manipulative suicide efforts. In the third study, Frances et al. (17), using a 1-hour unstructured clinical interview, obtained diagnostic information on 76 outpatients. They found that the 26 patients who met DSM-III criteria for borderline personality disorder were significantly more likely to report each of the eight DSM-III borderline personality disorder criteria than were SO control subjects with other axis II disorders. For the study that will be described here, we used the 22
summary
statements
of
the
Revised
Diagnostic
In-
terview for Borderline Patients (DIB-R), a semistructuned interview that can reliably distinguish clinically diagnosed borderline patients from patients with other axis II disorders (3 1), to assess the clinical features of borderline patients meeting DIB-R and DSM-III cnitena
and
control
subjects
with
other
This study builds on the design studies by incorporating all four tunes. First, diagnostic assessments clinical
diagnosis.
using
semistructured
liability.
Second,
Third,
assessments
inpatient
were
made
by
of demonstrated and
outpatient
reborder-
line patients and control subjects with other axis II disorders were studied. Fourth, control subjects meeting DSM-III criteria for a full range of axis II disorders were studied.
All subjects
were
in Belmont,
clinics
in the
munity
either
Mass.,
metropolitan
Counseling
at McLean
Boston
Center,
Counseling
Community
inpatients
or outpatients
Center,
Hos-
at one of three area
East
(Revere
Corn-
Boston/Winthrop
and Bunker
Hill
Health
Center in Charlestown). All patients at these facilities were considered eligible for inclusion if they 1) were between the ages of 18 and 40, 2) had normal or better intelligence,
3) had
no history
or current
symptoms
of
a serious organic condition or major psychotic disorden (schizophrenia on bipolar disorder), 4) had been substance free for at least 10 days, and 5) had been given definite axis II diagnoses by the admitting physician or referring clinician. Appropriate inpatients were
(D.L.C.),
who
outpatients pists. One
tient’s formed
screened
were of us
clinical consent
referred (M.C.Z.),
162
who
from
patient.
the phenomenological istening the DIB-R
each
and
by one
of us
admission notes; appropriate to the study by their therathen
status the
was
blind
obtained She
interview
a semistructured
the presence
of the
1 1 axis
that
II disorders
in DSM-III (32). who met both DIB and DSM-III criteria for personality disorder as assessed by the DIBDiagnostic Interview for Personality Disor-
dens
included
were
in the
borderline
group.
The
con-
trol group of subjects with other personality disorders comprised subjects who did not meet study criteria for borderline personality disorder but who did meet DSM-III criteria for some other type of axis II disorder as assessed by the Diagnostic Interview for Personality Disorders. Between-group comparisons involving categorical data were computed by using the chi-square statistic corrected for continuity; between-group comparisons involving continuous data (age and socioeconomic status)
were
computed
by using
analysis
post
test of pairwise
compar-
The Bonferroni correction for multiple was applied where applicable.
compar-
the Newman-Keuls isons. isons
hoc
of variance
and
RESULTS Overall,
to
the
written also
pa-
in-
evaluated
of each patient by adminDiagnostic Interview for
253
patients
were
interviewed
(148
inpa-
tients and 105 outpatients). Thirty of these patients did not meet DSM-III criteria for an axis II disorder and were excluded from further analyses. One hundred twenty of the remaining patients had borderline pensonality disorder; 103 had other axis II disorders. The mean±age of the patients with borderline personality disorder was 27.3 ±6.3 years; the mean age of the patients with other personality disorders was 28.1 ±6.3 The
mean
socioeconomic
status
of the
patients
with borderline personality disorder was 3.6± 1.3; that of the patients with other personality disorders was 3.6±1.1 (1=the highest status; S=the lowest). The 120 borderline patients did not differ significantly from the 103 patients with other personality disorders on the variables of race, marital status, age (see table 1), and socioeconomic background. However, the borderline
group
contained
a significantly
higher
per-
centage of women than the group of patients with other personality disorders (see table 1). Of the 103 control subjects with other personality disorders, 66 (64. 1 %) met DSM-III criteria for a “dnamatic” cluster disorder, 19 (18.4%) met DSMIII criteria for an “odd” cluster disorder, and 49 (47.6%) met DSM-III criteria for an “anxious” clusten
identified
diagnosis,
assesses
described Subjects borderline R and the
years. METHOD
pital
Disorders,
reliably
II disorders.
of the three earlier of the following feawere made blind to
instruments both
axis
Personality
disorder.
(These
percentages
add
up
to
more
than 100% because some patients met criteria for disorders in more than one cluster.) The most common disorder was antisocial personality disorder: 53 (51.5%) of the control subjects with other personality disorders met DSM-III criteria for this very-nearneighbor disorder. Table 2 shows the number of patients in each group who exhibited the clinical features detailed in the 22
Am
J
Psychiatry
147:2,
February
1990
ZANARINI,
TABLE 1. Demographic
CharaCteriStiCs
line Personality
and Other Axis II Disorders
Disorder
Patients With Other Axis II Disorders (N=103)
N
N
%
%
88 32
26.7
38 65
36.9 63.1
118
98.3
102
99.0
Other
733a
2
Marital status Ever married Never married asignificantly higher disorders
1.7
1
44 36.7 76 63.3 the percent of patients
than
pendency/ masochism, countertransference lationships.
were
44 59 with other
42.7 57.3 axis II
tunes-chronic
disorders were clinical features.
feelings
worthlessness/guilt,
of
judged to have Only four fea-
addition,
significantly
more
abuse/dependence,
sex-
Six
other
trol
subjects
at
less
O.OS-p